In April, when President Biden announced that American troops would leave Afghanistan, I wrote that there would probably never be a perfect moment to withdraw.
I still think that.
But four months later, watching what is happening across Afghanistan makes that argument considerably harder to say with confidence.
The United States is approaching the end of a twenty-year war. American and NATO forces are withdrawing. Bagram Air Base, once the centre of the American military presence in Afghanistan, has already been handed over.
But the war itself is not ending.
The Taliban is advancing.
Over the past several weeks, it has captured districts across the country, particularly in rural areas and along important border crossings. Now fighting is intensifying around major cities including Kandahar, Herat and Lashkar Gah.
Afghan government forces are fighting back. The United States has conducted air strikes in support of them.
But the direction of events is increasingly difficult to ignore.
America may be withdrawing from the war.
The Taliban appears to believe it can win it.
And that forces me to reconsider a question I thought I had mostly answered in April:
When does ending your participation in a war actually end your responsibility for what happens afterwards?
The map is changing quickly
The Taliban’s recent territorial gains have been extraordinary.
District after district has fallen.
Some Afghan soldiers have surrendered. Others have retreated. In some areas, local militias are mobilising alongside government forces.
The Taliban has also captured important border crossings, giving it greater control over movement and trade.
Until recently, many of its gains were principally rural.
That distinction matters because Afghanistan’s government still controls Kabul and other major population centres.
Holding countryside is not the same thing as governing the country.
But this week, the fighting has moved closer to important provincial capitals.
Lashkar Gah, the capital of Helmand province, is under severe pressure.
Fighting has intensified around Kandahar, Afghanistan’s second-largest city and historically one of the Taliban’s most important centres.
Herat in western Afghanistan is also facing a major Taliban assault.
This appears to be a new stage.
The Taliban is no longer simply demonstrating that the Afghan government cannot control every rural district.
It is testing whether the government can defend major cities without large numbers of foreign troops.
That is a much more important question.
Twenty years should have bought more than this
This is the part I find hardest to understand.
The United States and its allies have spent twenty years in Afghanistan.
Hundreds of thousands of Afghan soldiers and police have been trained.
Enormous amounts of money have been spent building the Afghan state and security forces.
Afghanistan has a government, parliament, army and air force.
And yet, as foreign troops leave, the Taliban is capturing territory at remarkable speed.
That raises an uncomfortable question.
What exactly was built during those twenty years?
This is not the same as saying nothing changed.
A generation of Afghans has grown up under a political system very different from Taliban rule.
Millions of girls have attended school.
Women have entered universities, journalism, politics and public life.
Cities have changed.
Institutions exist that did not exist in 2001.
Those achievements are real.
But institutions are only as strong as their ability to continue functioning when the foreign power helping construct them leaves.
If a state requires permanent external military support to survive, then perhaps we need to distinguish between building an institution and making it self-sustaining.
Afghanistan may be demonstrating the difference.
The Afghan army is stronger on paper
On paper, this should not necessarily be such an unequal fight.
Afghan security forces have more personnel than the Taliban.
They possess an air force.
They have heavy weapons and equipment supplied by the United States.
They have received years of training.
So why is the Taliban advancing?
Numbers do not capture everything that determines whether an army fights effectively.
Morale matters.
Leadership matters.
Logistics matter.
Whether soldiers are paid matters.
Whether ammunition reaches them matters.
Whether reinforcements arrive matters.
Whether they believe the government they are defending will still exist next month matters.
And perhaps most importantly, whether they believe the people beside them will continue fighting matters.
Military collapse can become self-reinforcing.
Imagine you are defending a district and hear that the district next to yours surrendered.
Then another falls.
You begin wondering whether reinforcements will ever arrive.
Your commander begins wondering the same thing.
At some point, surrender may stop looking like cowardice and start looking rational.
Then your surrender makes the calculation worse for the next unit.
Armies are organisations built partly from confidence.
Once confidence disappears, counting soldiers may tell us remarkably little.
The Taliban doesn’t have to be more popular
There is another mistake I think we should avoid.
Taliban military success does not automatically mean most Afghans support the Taliban.
Political preference and military control are not the same thing.
An armed organisation does not need majority support to capture territory.
It needs enough fighters, organisation, local networks and fear to make resistance difficult.
That distinction is important because otherwise every Taliban victory becomes interpreted as a referendum on twenty years of Afghan democracy.
It isn’t.
People can dislike the Taliban and still decide that fighting them is too dangerous.
Local leaders can negotiate surrender because they believe Kabul cannot protect them.
Soldiers can abandon positions without suddenly becoming ideological supporters of the people attacking them.
Control can come from consent.
It can also come from coercion.
Most governments use some combination of both.
Biden’s argument still has force
Despite everything happening this week, I do not think the answer is simply that Biden should cancel the withdrawal.
The strongest argument for leaving remains powerful.
If twenty years were insufficient to make Afghanistan politically stable, what would another year accomplish?
And if the answer is “another year would prevent the Taliban advancing”, then what happens at the end of that year?
Stay for another?
Then another?
At some point, temporary military support becomes permanent military presence.
Biden inherited a commitment negotiated by the Trump administration to withdraw American forces, although he changed the timetable.
He also inherited a war in which the original objective had changed dramatically.
America invaded Afghanistan after September 11 because al-Qaeda had used the country as a base while protected by the Taliban.
Osama bin Laden has been dead for a decade.
Al-Qaeda’s original infrastructure in Afghanistan was severely damaged.
The mission then expanded.
Defeat terrorism.
Defeat the Taliban.
Build Afghan security forces.
Build democratic institutions.
Protect women’s rights.
Create a stable government.
Prevent Afghanistan from again becoming a terrorist sanctuary.
Each objective may be defensible individually.
Together, they create a war with no obvious finishing line.
That is the problem Biden is trying to solve.
But “we cannot stay forever” avoids another question
There is a phrase used constantly in discussions about Afghanistan:
America cannot stay forever.
Obviously.
No military deployment should continue literally forever.
But that statement does not tell us whether this particular moment is the right moment to leave.
That distinction matters.
Suppose someone says:
“I cannot study forever.”
True.
It does not follow that they should leave the examination after twenty minutes.
The relevant question is whether additional time changes the likely outcome enough to justify its cost.
That is much harder.
If several thousand foreign troops could prevent the collapse of major Afghan cities at relatively limited cost, perhaps continued deployment is defensible.
But if those troops merely postpone the same collapse until whenever they eventually leave, continued deployment becomes much harder to justify.
The entire withdrawal debate depends on which of those descriptions is closer to reality.
And we don’t know yet.
This is partly a credibility problem
There is also a broader foreign-policy consequence.
America spent twenty years telling Afghans that it was helping construct a democratic political order.
People made decisions based on that promise.
Afghans worked with American forces.
Interpreters assisted foreign soldiers.
Women entered professions that Taliban rule had previously denied them.
Journalists criticised powerful people.
Politicians participated in democratic institutions.
If the Taliban returns to power, some of those people could face enormous danger precisely because they trusted the political order America helped create.
That creates a responsibility different from the responsibility to remain at war indefinitely.
America may reasonably decide that it cannot continue fighting.
But it cannot reasonably pretend that leaving makes every commitment disappear.
At minimum, Afghans who directly assisted American forces should not be abandoned because the immigration bureaucracy moves more slowly than the Taliban.
Responsibility can survive withdrawal.
Pakistan cannot be left out of this story
Afghanistan’s war also demonstrates why conflicts cannot always be understood inside national borders.
The Taliban has historically maintained important connections and sanctuary across the border in Pakistan.
Pakistan officially supports a negotiated settlement and denies supporting Taliban violence.
But the relationship has long been complicated.
From Pakistan’s strategic perspective, Afghanistan is connected to its rivalry with India.
Islamabad has historically feared an Afghan government closely aligned with New Delhi and has sought influence across its western border.
That means Afghanistan has not simply been an Afghan civil war plus an American intervention.
It has also existed inside a regional competition involving Pakistan, India, Iran, Russia, China and others.
Each possesses different interests.
This is why the phrase “Afghans should decide Afghanistan’s future” is morally appealing but strategically incomplete.
Afghans absolutely should decide their political future.
The problem is that neighbouring countries also possess the ability to influence which Afghans have enough power to make that decision.
Sovereignty is much easier to declare than to enforce.
The peace process looks increasingly detached from the battlefield
There are still negotiations between the Afghan government and Taliban representatives.
Diplomatically, everyone continues talking about a political settlement.
But negotiations depend partly on bargaining power.
Why would the Taliban make major concessions at a negotiating table if it believes its position on the battlefield is improving every week?
This is another connection between military power and diplomacy that I had underestimated.
Negotiations do not occur in a separate peaceful universe.
The battlefield changes what each side believes it can demand.
If I think waiting improves my position, I have less reason to compromise today.
That creates a dangerous incentive.
The Taliban may calculate that continuing the offensive will produce a better eventual agreement.
Or perhaps it believes it no longer needs an agreement at all.
If that calculation is correct, diplomatic appeals for compromise may accomplish very little.
What happens to Afghan women?
This may be the most morally difficult consequence of withdrawal.
The Taliban says it has changed since the government it operated between 1996 and 2001.
It says women’s rights will be respected within its interpretation of Islamic law.
But reports from areas under Taliban control have created serious fears about what that promise means in practice.
For twenty years, women’s rights have been one of the strongest moral arguments for the international presence in Afghanistan.
Now comes the uncomfortable question:
Is protecting rights in another country sufficient justification for indefinite military occupation?
My instinct is no.
If that becomes the standard, powerful countries would possess an almost unlimited justification for intervention anywhere rights are violated.
But saying military occupation cannot continue indefinitely does not make the consequences for Afghan women morally irrelevant.
There may simply be situations in which a country possesses enough power to prevent an injustice temporarily but not enough legitimate power to permanently reconstruct the political system producing it.
That is a horrible limitation.
It is still a limitation.
America may have confused dependence with success
Perhaps this is the larger lesson.
For years, progress in Afghanistan was partly measured while American troops were still present.
Afghan forces held cities.
Government ministries functioned.
Elections occurred.
Schools operated.
Those were genuine achievements.
But perhaps the real test was always what happened when the support structure was removed.
A bridge is not successfully built because it remains standing while scaffolding holds it up.
The test comes afterwards.
That does not mean the twenty-year intervention accomplished nothing.
It means we may have measured some of its accomplishments incorrectly.
Training an army is not the same as creating an army willing and able to operate independently.
Holding elections is not the same as creating political legitimacy.
Creating government institutions is not the same as creating a state capable of sustaining them.
Those differences are becoming visible now.
I still think America has to leave
This is where I arrive after watching this week.
I still think the United States eventually has to withdraw.
I still do not see a convincing explanation for how another five or ten years automatically produces the political settlement that twenty years failed to create.
But I am much less comfortable saying that than I was in April.
Because withdrawal is not simply ending a failed policy.
It is also removing something that has been preventing other failures from becoming visible.
Those failures now belong principally to Afghanistan’s political leaders and institutions.
But America helped construct those institutions.
It chose partners.
It spent money.
It designed military strategy.
It made promises.
That history creates obligations even if it does not create an obligation to fight forever.
So perhaps the question was never:
Should America stay or should America leave?
That makes the choice sound cleaner than it is.
The real question is:
How do you leave a war you cannot permanently win without abandoning the people who organised their lives around the political order you spent twenty years creating?
I don’t know the answer.
But watching Taliban forces move towards Kandahar, Herat and Lashkar Gah this week makes one thing increasingly difficult to believe.
Leaving Afghanistan and ending the consequences of the Afghan war are going to be two very different things.
